RE: Pango memory leak
- From: "Brandon M. Reynolds" <breynolds comtime com>
- To: 'Owen Taylor' <otaylor redhat com>, "Brandon M. Reynolds" <breynolds comtime com>
- Cc: "Gtk (E-mail)" <gtk-devel-list gnome org>
- Subject: RE: Pango memory leak
- Date: Fri, 12 Dec 2003 14:41:39 -0500
Ok, it looks like a false alarm. I did not profile the code long enough. I
was changing the font of a widget repeatedly, but Pango appears to do the
right thing. I think the memory consumption I am experiencing is due to
either gdk-directfb or directfb itself because when I run my program under
win32, the memory behaves nicely.
Brandon Reynolds mailto:bmr comtime com
Systems Engineer Ph: (330) 644-3059
Commercial Timesharing Inc. Fax: (330) 644-8110
> -----Original Message-----
> From: Owen Taylor [mailto:otaylor redhat com]
> Sent: Friday, December 12, 2003 11:09 AM
> To: Brandon M. Reynolds
> Cc: Gtk (E-mail)
> Subject: Re: Pango memory leak
>
>
> On Fri, 2003-12-12 at 10:36, Brandon M. Reynolds wrote:
> > I'm new to Gtk/Pango programming, so I could be wrong here.
> But I think
> > there is a memory leak in the function
> pango_layout_check_lines(). This
> > function declares a variable:
> > ParaBreakState state;
> > which contains a member called "glyphs" which gets memory
> allocated via
> > realloc(). It seems like pango_layout_check_lines() should
> free "glyphs"
> > (and others?) before it goes out of scope. Can anyone
> confirm/deny this?
> > I'm using Pango 1.3.0.
>
> The contents of "glyphs" is supposed to end up being moved into the
> various PangoLayoutRun->glyphs members that result from the layout
> process.
>
> Do you have details of the situation in which you think
> memory is being
> leaked?
>
> Regards,
> Owen
>
>
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]